Researchers deny mishandling sensitive information

OpenAI fired Jasmine Wang, Tomek Korbak, and Mikita Balesni last week. TechCrunch reports that on Thursday the three published an open letter to the company’s Safety and Security Committee, Safety Advisory Group, and Mission Advisory Council, denying that they mishandled sensitive information outside established procedures. The letter says internal and external communications around the firing have made former colleagues afraid to speak and operate in ways that, until last week, were part of ordinary work. The researchers write that people who work on safety see risks earlier and rely on close work with outside experts, and that the freedom to do so without fear, under defined internal procedures, is itself a safety mechanism. They also say conduct that was treated as normal a month ago is now grounds for dismissal, so remaining employees cannot tell where the line is. In the letter they deny leaking to The Information about newer models whose reasoning is harder to monitor, and they deny contacting outside parties beyond their jobs. The letter also addresses their response to the Hugging Face incident. It calls the incident and the investigation without precedent, and says internal policies were being written while the work was underway. According to the letter, Korbak believed that close communication with outside safety evaluators, in order to build trust, was still within the policies and norms of the time. It says Balesni worked internally on monitorability, checked with his reporting line and with board members and executives, and removed sensitive details from materials. This story does not describe how any agent left a test environment.

When TechCrunch published, OpenAI had not formally answered the letter. The company shared an internal memo attributed to a research leader. The memo says the decisions were not about raising safety concerns or speaking out, that the company does not fire employees for raising concerns, and that it values the three researchers’ safety work. A spokesperson separately told TechCrunch that an investigation found a pattern of misconduct, a clear violation of policies on handling research information, going beyond sharing information with an outside AI evaluation group. OpenAI did not tell TechCrunch which policies were violated, how the dismissals happened, or how it protects employees who raise safety concerns and work with outside evaluators. The Verge reports that on Friday OpenAI posted on X and stood by the firings. The company said an investigation found a significant breach of trust and a violation of clear policies on handling sensitive information, and that the decision was not about the three speaking out about the company or about safety. It also said the investigation found breaches beyond what the letter describes, and it gave no further details.

The two accounts do not match. The researchers say they believe they were fired for raising safety concerns and that they acted in line with the company’s mission and the working norms of the time. The company says the firings were not about speech, but about handling sensitive information. The investigation record is not public. This story sets the two accounts side by side and does not decide which one holds.